EXPERIMENTS WITH MODEL-DRIVEN DATA

ACQUISITION FOR CROWDSENSING

Phillip Dold 7/27/2013

Socially Relevant REU Program 2013

2

Air Pollution

You are concerned about air pollution in your city

Finding the causes of the pollution

Traditional Setup

Fixed Sensors

Crowdsensing Solution?

You could start a crowdsensing campaign

Recruit friends, family, and strangers

Collect particulates per million

3

Crowdsensing

Volunteers collect data with smartphones

Variety of sensors

Accelerometer

GPS

Light Sensor

Microphone

Cameras

4

Challenges of crowdsensing

Energy consumption

Sensors require energy

Communication is one of the biggest energy drains

Monetary Costs

Mobile data plans are not free nor “unlimited”

Both of these could decrease participation

5

Traditional Crowdsensing

Server /

Database

Collect data with

Users

Sends data to

6

Model-Driven Data Acquisition

Server

Model

Learns

Queries phones

Researchers

Queries

Queries phones

Users

Learning Phase

have

Query Phase Update Phase

Queries

7

Experimenting with Models

Implemented a simulator in Java that can be used

to experiment with models and implementations

Experimental Variables:

Degree of mobility

Density of network

Type of data

Length of learning phase

Evaluation of Metrics

Length of learning

Accuracy of model

Number of Updates

8

DBP (Derivative Based Predictions) [Raza 2012]

Simple Time Series

Model

Simpler Calculations

Less data needed

Expectation: Performance will drop with mobility

9

DrOPS (model-Driven Optimizations for Public Sensing )[Philipp 2013]

Multivariate Gaussian

Distribution Model

More Complex Calculations

More data needed

Expectation: Model will perform well, but will consume more energy than DBP

10

Experimental Setup

Simulator built in Java

Estimates Energy usage

Communication

Sensors

Datasets

Intel Lab

Lausanne Urban

Canopy Experiment

Mobility Traces

Cab spotting data from Crawdad

Intel sensor lab

11

Conclusions

Model-Driven Data Acquisition

Building a model rather than constantly sending data

It can help reduce communication

The simulator is still under development

Looking for additional data sets to use

12

Questions?

13

References

Philipp, D., Stachowiak, J., Alt, P., Durr, F., and Rothermel, K. DrOPS:

Model-Driven Optimization for Public Sensing Systems. In 2013 IEEE International

Conference on Pervasive Computing and Communications (PerCom) (PerCom 2013)

(San Diego, CA, USA, March 2013), IEEE Computer Society, pp. 1-8.

Raza, U., Camerra, A., Murphy, A. L., Palpanas, T., and Picco, G. P.

What does model-driven data acquisition really achieve in wireless sensor networks?

In Pervasive Computing and Communications (PerCom), 2012 IEEE International

Conference on (2012), IEEE, pp. 85-94.